PLENARY 4 – Friday 25

GO TO…

PLENARY 4 – Friday 25

Connecting bioacoustics to ecoacoustics to ecology. A new challenge for environmental monitoring and conservation through acoustics.

Gianni Pavan
Centro Interdisciplinare di Bioacustica e Ricerche Ambientali
Dipartimento di Scienze della Terra e dell’Ambiente
Università degli Studi di Pavia
Via Taramelli 24, 27100 Pavia, Italy
Email: gianni.pavan@unipv.it

Besides his anthropocentric view, Schafer was a pioneer in highlighting the need for soundscape research and management. In his book, The New Soundscape, Schafer (1969) documents rapid changes in soundscapes with the course of human civilization. However, Rachel Carson, in her book Silent Spring (1962), was the first to recognize the biophony as the natural expression of an ecosystem and to forecast its degradation in response to human impacts.

She expressed her concern for the massive use of chemicals and pesticides in agriculture and their impact on fauna and soil microfauna. She foresaw a silent world without the beautiful songs of birds, frogs, and insects killed by chemicals produced by industries and used in intensive agriculture, or these vocal animals disappearing because of dramatic habitat transformation imposed by industrial and urban development.

Following the development of Bioacoustics and Acoustic Ecology, now consolidated in the emerging field of Ecoacoustics, natural and anthropogenic sounds are recognized as ecological indicators and as essential drivers of several ecological processes in terrestrial and aquatic ecosystems. Now we consider the soundscape, in its strict description, as the result of human perception, and the acoustic environment, that exceeds the human hearing limits, extending into the realm of vibrations, infrasounds, and ultrasounds, even in water and below the threshold level of human perception. However, the term soundscape is often used in its broadest significance, largely exceeding human hearing and interpretation.

This discipline still requires continuous development, in theory, models, and methods, to connect Bioacoustics to Ecoacoustics and Ecology for environment management by providing robust ecological indicators in both the short and long time, in small areas as well as in wide ecological regions.

The use of ecoacoustic indicators, or indices, is still in its infancy. The emerging computational bioacoustics/ecoacoustics is slowly providing new tools to recognize specific sounds (species identification) and to generate global indices to robustly describe ecosystem by identifying sonotopes and acoustic communities.

However, a widely accepted protocol for data acquisition and processing is still undefined, as well as it is not clear how to connect some indices to ecosystem structures and processes.

All these advances support the development of the soundscape conservation issue and the study of the noise pollution effects on the health of ecosystems and of human beings.

The acoustic quality of the environment in wild and remote ecosystems, and in human-impacted areas, represents a value to be preserved and improved for the well-being of animals and humans. To conclude, we need to establish largely accepted and robust protocols to describe the evolution of sonosystems locally and globally to support nature conservation.

O43- Singing in a noisy ocean: can male humpback whales cope with natural and anthropogenic noise?

  • Girola, E., School of Veterinary Science, University of Queensland, Brisbane, Australia
  • Noad, M., School of Veterinary Science, University of Queensland, Brisbane, Australia
  • Dunlop, R., School of Veterinary Science, University of Queensland, Brisbane, Australia

The soundscape in the shallow waters off Eastern Australia is characterised by a variety of natural and anthropogenic sources. Amongst these, wind generated noise and vessel noise are constantly present, although their levels are variable. In the winter months, humpback whales travel through this area during their annual migration between feeding and breeding grounds. While in transit, the males produce complex series of vocalizations, known as songs, believed to be a reproductive display. Underwater noise can mask whale songs limiting their range and reducing signal clarity. This project investigates whether singing whales modify the characteristics of their vocalizations to compensate for changes in the soundscape to maintain efficient communications.

Underwater noise and humpback whale songs were recorded using a fixed hydrophone array deployed off the Queensland coast. Due to natural variability, the recording period provided a range of wind noise levels. At the same time, a series of controlled experiments was carried out to expose nearby whales to variable levels of vessel noise. The recordings were analysed to look for correlations between the characteristics of the songs and noise levels.

Results showed that when the soundscape was dominated by the wind, singing whales compensated for variable levels of noise by adjusting the source levels of their songs, but they did not modify the spectral and temporal parameters. During the experiments, when vessel noise was dominating the soundscape, there was no correlation between song characteristics and vessel noise levels; however, the singers adjusted their source levels to the underlying wind noise.

This study found that male humpback whales do not sing as loudly as possible to communicate their fitness or to maximize their communication range and that, although they do not modify their sounds to compensate for anthropogenic noise, they regulate their songs based on the natural components of the soundscape.

O44- Developing a fine-scale soundscape spatial model for measuring biodiversity in a semi-arid landscape 

Marina D. A. Scarpelli a, David Tucker a, Susan Fuller b and Paul Roe a

aQUT Ecoacoustics Research Group, School of Computer Science, Science and Engineering Faculty, Queensland University of Technology, Australia

bQueensland University of Technology, School of Biology and Environmental Science, Science and Engineering Faculty, Brisbane, Queensland 4001, Australia

Vegetation composition and structure and water availability are key determinants of the faunal communities found within semi-arid and arid ecosystem regions. This research will examine how soundscapes vary across spatially distinct but proximate vegetation communities in semi-arid Australia and identify the main drivers of this variation. Acoustic recorders were deployed at Bowra Wildlife Sanctuary in semi-arid western Queensland, Australia, along a transect at 200m intervals for 24 hours in August and October 2019. Bowra Wildlife Sanctuary has a mostly flat (low elevation) landform, and is dominated by Acacia woodlands, tussock grasslands, and Eucalyptus coolabah woodlands fringing ephemeral creek lines. Deployment points were selected using two main criteria: (1) to maximise the captured variability across multiple vegetation communities, and (2) to consider proximity to water, which is a valuable resource in semi-arid regions. Local weather data were also collected, as well as surrounding vegetation community and bird species richness data.
Acoustic recordings were analysed using a suite of acoustic indices and clustering techniques. Preliminary data will be presented showing how cluster data can be combined with environmental variables to build a spatial model indicating the locations and features driving soundscape variability. Maps will be produced showing the soundscape at different points, and the soundscape patterns at different times of the day.

O45- What does Atlantic Forest soundscapes can tell us about landscape?

 

Marina D. A. SCARPELLI1, a, Milton Cezar RIBEIRO2, Camila P. TEIXEIRA3

1 Environmental Analysis and Modelling Graduation Program – Federal University of Minas Gerais (UFMG) – Geosciences Institute, Minas Gerais, Brazil

2 Spatial Ecology and Conservation Lab (LEEC) – Sao Paulo State University (UNESP) – Biosciences Institute/Rio Claro, São Paulo, Brazil

3 UNIFEMM – Sete Lagoas University Centre – Minas Gerais, Brazil

a QUT Ecoacoustics Research Group, Science and Engineering Faculty, Queensland University of Technology, Australia

The ecoacoustics approach for environmental recordings analysis is used to understand and identify big patterns related to different sound sources, like animals, humans and the environment itself. Sounds can vary according to several features that can be on its nearby surroundings or in a distance, therefore they are very much reliant on scale. Since humans are changing the environment so much and we cannot account for all those changes in the same speed as they happen, we need fast evaluation tools, such as remote sensing and acoustic monitoring (the equivalent of spatial remote sensing for sounds). Considering that the scale of effect was never measured for soundscapes before, we tested different buffers sizes to assess at what scale different acoustic indices were responsive to. Also, we tested how natural vegetation cover influenced different acoustic indices. We recorded environmental sounds in Atlantic Forest areas during three months on the rainy season. Then we calculated different acoustic indices and the percentage of natural vegetation cover in different scales. Our hypotheses were corroborated by our results: different indices respond to different scales and their medians varied according to the amount of vegetation cover on the surroundings. More studies are needed in less fragmented areas, to test indices behaviour in a continuum, but we consider this work an important start to understand acoustic indices behaviour in tropical areas, especially in such degraded area as Atlantic Forest.

O46- What can Mathematics teach us about the spectrogram?

  • Listanti, V., School of Mathematics and Statistics, Faculty of Science, Victoria University of Wellington, Wellington, New Zealand
  • Marsland, S., School of Mathematics and Statistics, Faculty of Science, Victoria University of Wellington, Wellington, New Zealand
  • Priyadarshani, N., School of Mathematics and Statistics, Faculty of Science, Victoria University of Wellington, Wellington, New Zealand

Being easy and quick to read, the spectrogram is central to data analysis in bioacoustics. However, we rarely think about how to take advantage of its properties. When computing a spectrogram, we use Fourier Transform to turn a sound wave, a function of pressure and time, into a picture of energy in time-frequency plane. This picture is usually used to characterise and classify animal sounds or to design an automatic recogniser.

We will discuss how a deeper understanding of the spectrogram can improve the analysis of real word data.

As an example, we will present our work on the echolocation of the two New Zealand bats species: Lesser short-tailed, Mystacina tuberculata, and Long-tailed, Chalinolobus tuberculatus. We worked with data collected using the recorders at the disposal of New Zealand Department of Conservation (DOC). These devices record suspected bat pass into a compressed version of the spectrogram which retain sufficient information to recognise the echolocation calls.

We will demonstrate a Convolutional Neural Network algorithm for the automatic detection and classification of these calls. Moreover, we will discuss how we can retrieve the lost original sound with spectrogram inversion.

This technique is based on the mathematical definition of the Fourier Transform, which is the most exploited time-frequency representation (TFR) of signals. However, it is not flawless. Due to the uncertainty principle, there are limits in the time-frequency resolution, that cannot be corrected by its hidden parameters (e.g. window length, type or overlap).

We will briefly introduce other mathematical tools that can be used as TFRs, discussing how to compare them and what can be the future developments.

O47- Exploring phenological asynchrony between avian diversity and vegetation in temperate deciduous forests through bioacoustic monitoring and camera trapping.

  • Barnett, R., Institute of Biodiversity, Animal Health and Comparative Medicine. University of Glasgow, Glasgow, UK
  • McGregor, A., Institute of Biodiversity, Animal Health and Comparative Medicine and School of Life Sciences. University of Glasgow, Glasgow, UK
  • White, S., Institute of Biodiversity, Animal Health and Comparative Medicine and School of Life Sciences. University of Glasgow, Glasgow, UK

Climate change has had significant effects on the phenology of spring-time vegetative growth and bird reproduction, however, increasing evidence exists that reproduction in long-distance migratory bird species is not keeping pace with the emergence of vegetation in temperate habitats. The aim of this study was to explore whether such asynchrony could be occurring in Scottish deciduous forests using acoustic monitoring and camera traps to track the phenologies of migrant birds and vegetation growth from January to July 2019. A series of images was obtained using time-lapse settings of camera traps to determine a ‘greenness score’ as vegetation emerged from winter dormancy. Simultaneous bioacoustic monitoring was used to estimate the arrival date of eleven species of singing passerine and the difference between green-up and arrival was calculated to measure the phonological interval. Acoustic indices were also calculated and compared against avian and invertebrate diversity, measured through mist-netting and trapping studies respectively, to determine whether acoustic diversity could be an effective measure of overall biodiversity before being compared with greenness index. The Bioacoustic Index was determined to be the most promising index for estimating avian diversity. The Bioacoustic Index and Normalised Difference Soundscape Index also showed strong positive associations with photoperiod, likely as a consequence of increased daylength triggering mating behaviour and increased bird song in the soundscape. Short-distance migrant species tended to arrive earlier than long-distance species, in keeping with trends observed elsewhere. In summary, there did appear to be phenological asynchrony between migrant birds and vegetation within these forests, however more work is needed to connect bioacoustic indices with measured avian and invertebrate diversity.

O48- The acoustic diversity and complexity of the Maromizaha Forest Reserve, Madagascar

  • Marco Gamba, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Valeria Torti, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Chiara De Gregorio, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Daria Valente, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Longo Miaretsoa, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Rose Marie Randrianarison, Group d’Étude et de Recherche sur les Primates de
  • Madagascar, Antananarivo, Madagascar
  • Teresa Raimondi, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Olivier Friard, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Livio Favaro, Dept of Life Sciences and Systems Biology, University of Torino, Italy
  • Cristina Giacoma, Dept of Life Sciences and Systems Biology, University of Torino, Italy

The diversity of acoustic sensory systems and the complexity of acoustic stimuli in the environment shape how animals behave and interact with their surroundings. How the acoustic environment drives behavioral variation that can operate at short time scales can be ideally investigated in Madagascar rainforests. They represent a peculiar situation in which the human impact on managed forest areas led to the identification of different regions of increasing protection enforcement (e.g., touristic, research, and pristine). We focused our research on the ecoacoustics of different areas belonging to the touristic, research, and pristine regions.

Our results showed that soundscape patterns changed significantly according to the different protection measures. The Acoustic Diversity Index of the pristine forest areas differed significantly from that of the research and the touristic regions, which instead did not show substantial differences between them. The Entropy H Index and the Acoustic Complexity Index showed that differences were mainly involving the research area versus the other two regions.

We conclude that vegetation patterns strongly influence soundscape patterns in montane Malagasy rainforest and that this may result in behavioral adaptations of the critically endangered species living in these areas. In a scenario in which human impact determines an accelerated biodiversity loss, it will be essential to investigate how these indices may vary over time.

O49- At-sea acoustic tracking of seabirds: exploring the soundscapes of highly mobile predators during foraging.

  • Cerqueira, A. S., King’s College London and ZSL Institute of Zoology, London, United Kingdom.
  • Dawson, T. P., King’s College, London, United Kingdom.
  • Freeman, R., ZSL Institute of Zoology, London, United Kingdom.
  • Phillips, R. A., British Antarctic Survey, Cambridge, United Kingdom.

Seabirds are very well-studied in comparison to other groups of animals with equivalent role in the marine environment; however, their behaviour at sea remains difficult to observe and understand, given that these animals are highly mobile, wide-ranging, can spend long periods of time in the open ocean and are often elusive. Recent advancements in the development of biologging technologies have led to the creation of improved, miniaturised sound recording devices that can be deployed directly on animals, enabling scientists to sample their soundscapes in a variety of settings, remotely.

In this study, we combine the use of miniature bird-borne audio recorders with GPS trackers to explore the soundscapes of two albatross species breeding at Bird Island, South Georgia, to investigate their behaviour and environmental context during foraging at sea. Our findings show that at-sea seabird sound recordings allow the precise characterisation of activities they engage in, helping to refine animal movement data, and that events detected through sound recordings are associated with the duration of seabirds foraging activities. We also found that seabird sound recordings permit pinpointing of specific locations at sea where birds congregate.

Seabird soundscapes are a repository of useful information waiting to be explored, with potential to unveil new insights into seabird social behaviour and interactions with their environment. By developing our understanding of seabird activities during foraging and mapping the locations where specific behaviours at sea occur, we can inform policy to support seabird protection where and when it is most needed.

O50- POSTER- Acoustic monitoring of wetland habitats in dry regions (Kuwait): bird community dynamics related to migration

  • Sarah Obaid, Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France &
  • Frédéric Jiguet, Centre d’Écologie et des Sciences de la Conservation (CESCO), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, 57 rue Cuvier, 75005 Paris, France
  • Jérôme Sueur, Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France

Wetland bird communities are particularly complex due to a mix between resident and migratory species. The arrival and departure of non-resident species induce important dynamics in terms of species richness and assemblage composition. These dynamics make the monitoring of wetlands rather complex.

Wetlands in Kuwait play a very important role for biodiversity as a shelter for many migrant birds, a breeding habitat for residents and a non-breeding habitat for wintering birds, a foraging resource for many species. Artificial and constructed wetlands have helped with restoring several species that avoided coming to Kuwait. With the availability of shelters and food that these new wetlands offer, some species return back to winter or/and breed again.

So far, terrestrial eco-acoustic studies have mainly focused on tropical and temperate habitats but very rarely on open habitats such as wetlands. The main objective of this project is to use the eco-acoustic approach to monitor wetland sites in Kuwait, in order to better understand the local ecological dynamics and to help nature preservation.

We plan to describe and monitor the wetland bird community dynamics of Kuwait at several sites. More specifically, we will comprehensively assess the bird population, phenology, turnover, and richness in Kuwait (objective 1), estimate of Alpha diversity using acoustic diversity indices (objective 2), and test the possibility to use acoustics to estimate the phenology of bird migration (objective 3).

O51- Semiochemicals, semiophysicals and their integration for the development of innovative multi-modal systems for agricultural pests’ monitoring and control

  • Nieri, R., Department of Civil, Environmental and Mechanical Engineering, University of Trento, Italy; Research and Innovation Center, Fondazione Edmund Mach, 38010 San Michele all’Adige, TN, Italy
  • Anfora, G., Center for Agriculture, Food and Environment (C3A), University of Trento, 38010 San Michele all’Adige, TN, Italy; Research and Innovation Center, Fondazione Edmund Mach, 38010 San Michele all’Adige, TN, Italy.
  • Mazzoni, V., Research and Innovation Center, Fondazione Edmund Mach, 38010 San Michele all’Adige, TN, Ital
  • Rossi Stacconi, M. V., Research and Innovation Center, Fondazione Edmund Mach, 38010 San Michele all’Adige, TN, Italy.

Herbivorous insects are primarily thought to depend on olfaction and taste for their intra- and interspecific communication. Thus, semiochemicals (i.e., chemical signals) have been widely studied for pest management applications. However, given that pest behavior does not rely on just one communication modality, stimuli of physical nature, such as light, sounds and vibrations, can also be used to manipulate insect-insect or insect-plant interactions. Moreover, stimuli of different natures can be combined in a multi-modal pest management program to increase the overall efficacy. Besides the widespread use of both chemical and physical signals in multimodal insect communication, the integration of stimuli has hardly been implemented for hardly any crop. This review introduces the term semiophysicals as opposed to semiochemicals and focuses on how pest behavior can be manipulated by discussing three main approaches; i) manipulation of pest orientation through attractive/repellent stimuli, ii) inhibition or promotion of specific pest behaviors and iii) interference with intraspecific communication through disruptive stimuli. For each approach, we provide examples of use of both semiochemicals and semiophysicals. Lastly, we describe the case study of the vineyard agroecosystem in the Trento province, where a multi-pest management program has been successfully developed, and we discuss future perspectives.

O52- Elucidating the acoustic dynamics of Indirana chiravasi with special emphasis on temperature and humidity

Shashank Borkar*,1 and Nikhil Modak2

1Department of Zoology, Modern College of Arts Science and Commerce, Shivaji Nagar, Pune

2Bombay Natural History Society, Mumbai

*shashankborkar57@gmail.com

Anuran amphibians use acoustic clues of their surroundings (extrinsic factors) to make their signals meaningful and decodable. Also, their morphology and anatomy (intrinsic factors) play an important role in responding to the environmental clues. Currently, we are investigating the effects of temperature and humidity of surroundings on the acoustic dynamics of Western Ghats endemic species Indirana chiravasi, from its type locality and role of body length, weight, and temperature of individual in responding to changes in surroundings. We realize the response of the species through call rate, average dominant frequency, atmospheric attenuation and speed of sound. The calls of fifty individuals which were recorded manually and analyzed for their temporal and spectral properties in Raven 2.0. We used Partial Least Square Regression (PLSR) approach to understand the interplay between dependent and predictor variables. The significance of the results was determined after Bonferroni sequential correction. Our results bolster known physical phenomenon that the speed of sound increases with increase in temperature (p < 0.0001) while the same decreases with increase in humidity (p < 0.005). The atmospheric attenuation showed strong positive correlation to average dominant frequency (p < 0.0001). Also, the average dominant frequency is strongly negatively correlated with body size (SVL) (p < 0.005). The call rate did not show any correlation with variations in temperature and humidity. To check if Temperature-Size rule fits to our model we checked dependency of SVL against Temperature. The SVL showed strong negative correlation (p < 0.0001) with temperature which implies the fact that this species follows the temperature-size rule. However, this fact is worrying, as increase in temperature will favor smaller individuals which can produce higher frequencies than larger individuals which ultimately are attenuated more than lower frequencies thus taking a toll on acoustic facilitation of the species. These results become pertinent when we are expecting increase in average global temperature which may become acoustically detrimental to this species.

Key Words: Biodiversity Hotspot, Climate Change, Temperature-Size Rule, Ecoacoustics

O53- SOUND EARTH LEGACY

  • Andrea Lamount
  • Sound Earth Legacy

Summary; Sound Earth Legacy is a non-profit organization aiming to preserve the sounds of the earth and supporting pioneer scientific and environmental projects through sound and music. Discover their current projects in the presentation. The organization is open to new collaborations

O54- Ecotremology – new insights into hidden ecosystems: a pioneer study in meadow habitats

  • Šturm Rok, National Institute of Biology, Večna pot 111, SI – 1000 Ljubljana, Slovenia
  • Juan José López Díez, National Institute of Biology, Večna pot 111, SI – 1000 Ljubljana, Slovenia
  • Behare Rexhepi, National Institute of Biology, Večna pot 111, SI – 1000 Ljubljana, Slovenia
  • Sueur Jérôme, Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France
  • Virant-Doberlet Meta, National Institute of Biology, Večna pot 111, SI – 1000 Ljubljana, Slovenia

In 2016, biotremology – study of mechanical communication by surface-borne waves, was defined as an independent scientific discipline. The transmission media (air vs solids) impose different constraints on signal propagation so organisms using different media follow different evolutionary paths.

In our work, we focused on meadow habitats from Slovenia. The vibroscape structure and dynamics were analyzed at different temporal scales, from diel variation to seasonal changes. Substrate vibrations were recorded from plant stems using laser vibrometers. Vibrational events were manually assigned to vibrational signal types (VST) according to their temporal and spectral characteristics. Vibrational signals are species- and sex-specific. Due to the lack of verified publicly available library of vibrational signals, only 13 out of 60 registered VSTs could be identified to species level.

The highest abundance of VSTs was observed at the beginning of July, when the vibrational community included 14 species with three VSTs dominating. The overlap of these signals in time and frequency domain was significantly smaller than it would be by chance. This reveal for the first time a partitioning of the vibroscape, which suggests existence of species interaction for communication channel. Diel variation in vibrational signaling activity was correlated with ambient temperature (Pearson r = 0,7). Wind provided nearly constant background vibrational noise but higher wind velocities (> 0,8m/s) reduced the amount of biological component of vibroscape. Results also revealed differences in vibroscape composition between hairy sedge (Carex hirta) and hedge bedstraw (Galium mollugo) plants growing on the same meadow. The former included a higher VST richness and higher abundance turnover among individual plants, which may be attributed to plant geometry and host specific plant-animal interaction.

In summary, the vibroscape of a meadow revealed a rich and complex vibratory world which is not directly accessible to humans. Vibrational signaling is the most common form of mechanical communication, particularly common in arthropods, and as such, surface-borne mechanical waves are commonly present in environment and offer readily available and reliable source of information on ecological processes in hidden ecosystems.

O55- Ecoacoustics as a novel tool for assessing pond restoration success: Results of a pilot study 

Jack Greenalgh
University of Bristol, UK
24 Tyndall Ave, Bristol BS8 1TQ, Bristol

  1. Ecoacoustics is increasingly being used to monitor species populations and to estimate biodiversity in marine ecosystems, but the underwater soundscapes of freshwater environments remain largely unexplored in this respect. Few studies exist concerning the acoustic diversity of ponds, but because aquatic plants and many arthropods such as Coleoptera and Hemiptera are known to produce sound, there is potential to use ecoacoustic techniques to monitor changes in biodiversity and conservation value.
  2. This pilot study compares the underwater soundscapes of recently restored open-canopy ponds and unmanaged highly terrestrialized ponds situated in an arable agricultural landscape of North Norfolk, UK, in order to assess the benefits of farmland pond restoration.
  3. Daytime sound recordings were made for 10 min in each pond and analysed primarily for arthropod stridulations. In addition, six commonly used acoustic indices were calculated to assess the soundscape biodiversity between the unmanaged and the restored ponds. The stridulations of three diving beetle species (Dytiscidae) were recorded in tank studies to assess the potential for individual species recognition from underwater sound capture.
  4. Sound-type richness and abundance, as estimated by visually and aurally identifying arthropod stridulation from spectrograms, were significantly higher in the restored open-canopy ponds compared with the unmanaged terrestrialized ponds. In addition, the acoustic indices ‘acoustic complexity’ and ‘biodiversity index’ were significantly higher in restored open-canopy ponds than in unmanaged terrestrialized ponds.
  5. The three dytiscid water beetle species recorded in a tank were found to produce distinctive and recognizable sounds, indicating potential to create an audio reference library that could be used for automatic acoustic monitoring of freshwater arthropods.
  6. Pond soundscapes are rich in biological information and this study suggests that, with further development, automated passive ecoacoustic monitoring could be an effective non-invasive technique for assessing pond conservation value and pond restoration and management success.

O56- Visualization and Data Science Advances in Soundscape Ecology

  • Rosane Minghim, University College Cork, Ireland
  • Liz Maribel Huancapaza Hilasaca, University of S˜ao Paulo, Brazil
  • Fabio Felix Dias, University of S˜ao Paulo, Brazil
  • Moacir Ponti, University of S˜ao Paulo, Brazil

The Science and Technology of soundscape ecology is bound to have a growing impact in the way natural environments are studied, managed and monitored. The amount and complexity of the data collected demand the development of better computational strategies that can evolve to automatic and semi-automatic data processing. At the current stage of data handling in soundscape ecology, the user must be in the center of the discovery process and there is the need for focused effort in developing and adapting data and visual analytics strategies to meet the challenge.
Some of the challenges regarding the development of proper tools are:

  1. the scale: the same source (recordings) is meant to support a large number of studies, from evaluation of diversity to recovering and estimating presence and abundance of particular species;
  2. labelling: for many different applications, accelerating and automating labelling processes is paramount;
    and
  3. feature analysis: from extracted features to neural network embeddings it is necessary to understand and explain to final users what they are meant to represent.

In this presentation we report on latest results and current developments in tackling these issues for the case of soundscape ecology. We describe tools for landscape and species discrimination by means of both extracted features and neural network embeddings. We reflect on the application of multidimensional visualisation and active learning strategies to support the labelling processes, and we illustrate the application of feature analysis strategies and their role in understanding segregation of soundscapes.

O57- Can we estimate marine biodiversity using sound recordings? Application of acoustic indices at Mozambique Island coral reefs.

  • Manuel Vieira, Departamento de Biologia Animal and cE3c – Centre for Ecology, Evolution and Environmental Changes, Faculdade de Ciências, Universidade de Lisboa, Portugal; MARE – Marine and Environmental Sciences Centre, ISPA-Instituto Universitário, Lisboa.
  • Ricardo Duarte, Universidade Eduardo Mondlane, Maputo, Mozambique
  • Paulo J. Fonseca, Departamento de Biologia Animal and cE3c – Centre for Ecology, Evolution and Environmental Changes; Faculdade de Ciências, Universidade de Lisboa, Portugal.

Passive acoustic monitoring (PAM) is emerging as an easy, efficient and non-invasive tool to evaluate and monitor the marine environments. Interest in surveying techniques based on acoustic indices is growing among marine ecologists. However, the relation between acoustic indices and biodiversity is not clear for marine habitats. In this study we have tested the potential of three acoustic indices, namely the sound pressure level, the acoustic complexity index, and the temporal entropy, using 24h of acoustic continuous recordings on several sampling points at Mozambique Island coral reefs, to evaluate fish biodiversity. Fish species richness was visually surveyed at each recording place. Sound pressure level was used to segment the recording in percentiles of energy. Indices were averaged for three frequencies bandwidths: 10-1000 Hz, 1000-2000 Hz and full bandwidth up to 22050 Hz. The acoustic complexity index on the frequency bandwidth from 1000-2000 Hz was positively related with fish diversity (R2= 0.4). The average sound pressure level showed also a positive relation with fish diversity, especially if excluding the 0.01 and 0.99 percentiles. Temporal entropy showed no relations with fish diversity. These results suggest that bioacoustics indices can be used as a cost-effective tool to monitor marine environments, but further investigation is needed to interpret the indices values and standardize its use across habitats.

O58- Bats on Waterways inside a City: Bioacustic Research in Brescia (Lombardy, Italy).

  • Vincenzo Ferri, Christiana Soccini,
  • Centro Studi Naturalistici Arcadia, via Valverde 4, 01016 Tarquinia, Italy.

Urban waterways are important habitat for wildlife including bats: along them there may be important roosting sites, as crevices in stone work of old and modern bridges and they are also important foraging areas for these threatened mammals.

In the course of a broader research on distribution and situation of Chiroptera populations, bioacoustic surveys were carried out to record attendance and foraging activity of bats on waterways inside the city of Brescia. In the urban context rivers, streams and drainage ditches represent the only line of ecological permeability and connection between residual elements of naturalness in this territory.

Most of accessible sections of Garza and Garzetta torrents, of Mella river and of Naviglio Grande, N. Cerca and N. Inferiore were surveyed. Some sections, from historical times with a completely underground course, have been monitored at the beginning and end of the underground section. Searches were carried out between months of May and September of 2019, in 25 stations, with 13 bioacoustic detection sessions and 75 locations of 4 hours, for a total of 300 hours of recording, using 6 audio-ultrasonic microphones: 2 Ultramic 384K and 4 Ultramic 384K BLE (Dodotronic Srl, Castel Gandolfo, Italy), with sampling frequency of 384 kHz, on SD memory cards in wav files. Bioacoustic analysis of 1113 recorded bat passes (BP) allowed identification of 7 Chiroptera species, including Myotis daubentonii and Miniopterus schreibersii; Pipistrellus kuhlii (57.32% of the BP), Hypsugo savii (21.02%) and Pipistrellus pipistrellus (16.08%) were most common and ubiquitous species. General average Activity Index detected was 3.71 BP / h, with a maximum of 9.75 BP / h. Simultaneous use of ultrasonic detectors at the ends of investigated underground sections of Garza torrent has proved that bats do not roost there, but that they go on uncovered sections to feed flying invertebrates (Diptera Culicidae, Tipulidae, Chironomidae and Simulidae) in emergency from water or flying towards surrounding public lights.

O59-Listening to climate change: a framework to forecast species spatial and phenological shifts using acoustic monitoring

  • Llusia, D., Terrestrial Ecology Group, Science Faculty, Universidad Autónoma de Madrid, Madrid, Spain
  • Desjonquères, C., Terrestrial Ecology Group, Science Faculty, Universidad Autónoma de Madrid, Madrid, Spain
  • Márquez, R., Biodiversity and Evolutionary Biology Department, Museo Nacional de Ciencias Naturales, Madrid, Spain
  • Beltrán, J.F., Department of Zoology, Universidad de Sevilla, Sevilla, Spain
  • De Marco, P., Instituto de Ciências Biológicas, Universidade Federal de Goiás, Goiânia, Brazil
  • Villén-Pérez, S., Global Change Ecology and Evolution group, Department of Life Sciences, University of Alcalá, Alcalá de Henares, Spain

Ecoacoustics offers a promising basis to explore the capacity of species to deal with climate change. The use of eco-physiological data has been suggested to develop mechanistic models as a key tool in climate change research, although there is still a vast lack of information. Now passive acoustic monitoring may assist in filling this gap. Recording species calling behaviour over their distribution ranges allow us to estimate the climatic breadth of calling performance and provides a baseline for predictive models. Here, we introduce a general and transferable methodological framework to combine acoustic and biogeographic information aiming at forecasting both phenological and geographical shifts of calling behaviour in response to climate change. We develop two complementary approaches: a boundary model that is based on the recorded climatic breadth of calling and a regression model that fits acoustic activity and climate using generalized linear models. Using a study case based on a continuous acoustic monitoring of populations located at the thermal extremes of the species range, we test these two models, predicting present and future climatic suitability for calling behaviour of Iberian frogs. Both models have high and similar prediction accuracy, and forecast a phenological advance of acoustic activity and a northward and westward shift of suitable climatic conditions for calling behaviour by 2050 under RCP8.5 climatic scenario. Our results suggest that climate change could drive phenological and geographical shifts in the calling activity of the studied species. Our study demonstrates how acoustic monitoring techniques represent a valuable opportunity for the field of climate change research.

O60-  Biogeography of fish sounds: Acoustic communities are tuned by the habitat

Lucia Di Iorio1,2, Manon Audax1, Philippe Lenfant2, Julie Deter3,4, Cédric Gervaise1, Pierre Boissery5

1 CHORUS Institute, 115 rue des Alliés, 38100 Grenoble, France

2 CEFREM, Université de Perpignan, 52 avenue Paul Alduy, 66860 Perpignan

3 Andromède Océanologie, 7 place Cassan – Carnon plage, 34130 Mauguio, France

4 MARBEC, Univ Montpellier, CNRS, IFREMER, IRD, Place Eugène Bataillon, 34095 Montpellier, France

5 Agence de l’Eau Rhône Méditerranée Corse, Imm Le Noailles 62 La Cannebière, 13001 Marseille, France

Monitoring marine biodiversity and understanding their drivers across geographic scales is essential to preserve ecosystems functions and associated services. Monitoring marine habitats, their responses to environmental pressures or management actions is often challenging, in particular at large scales. Ecoacoustics is a promising avenue, yet the drivers of acoustic community composition remain unknown, as well as to which extent acoustic biodiversity can reflect environmental status and the effectiveness of protection measures. Here we unveil the biogeography of fish sounds of 27 coralligenous reefs across the North-Western Mediterranean, covering 2000km and three regions. We also applied community ecology principles to the fish sounds of a marine protected area (MPA) with different protection levels to evaluate how habitat and protection levels shape acoustic fish communities. We found 28 putative coralligenous fish sound types, which is twice as much as recorded in other Mediterranean habitats making coralligenous reefs an acoustic hotspot. 40% of these sounds are not found in other coastal habitats, thus likely specific to coralligenous reefs. Acoustic diversity differed between geographical regions. Ubiquitous sound types were identified, among these, sounds from top-predator species, and others more specifically related to the presence of ecosystem engineers (red coral, gorgonians), which are key players in maintaining habitat function. Depth, percent of coralligenous outcrops and habitat condition were the main determinants of acoustic community composition indicating that acoustic fish communities are related to benthic assemblages. We also found that acoustic community composition and diversity were mainly driven by the MPA’s protection levels. Fully protected sites were more diverse and hosted more sounds from a few high-level predators than partially protected sites. This work reveals that acoustic biodiversity can contribute to habitat biogeography, depict habitat conditions, be indicative of protection levels, and infer information on ecosystem functioning. This is highly relevant for conservation and habitat monitoring.

O61-POSTER-  Ecoacoustics in Support of Quiet Area Ecological Connectivity Plans

  • Aggelos Tsaligopoulos
    Acoustic Ecology Laboratory, Department of Environment, University of the Aegean, Mytilene, Lesvos, Greece, email: tsaligopoulos@env.aegean.gr
  • Stella Kyvelou
    Department of Economics and Regional Development, School of Sciences of Economics and Public Administration, Panteion University of Social and Political Sciences, Athens, Greece, email: kyvelou@panteion.gr
  • Aimilia Karapostoli
    School of Architectural Engineering, Democritus University of Thrace, Xanthi, Greece, email: akarapos@arch.duth.gr
  • Antonella Radicchi
    Technical University Berlin, Institute of City and Regional Planning, Berlin, Germany, email: antonella.radicchi@tu-berlin.de
  • Chris Economou
    Acoustic Ecology Laboratory, Department of Environment, University of the Aegean, Mytilene, Lesvos, Greece, email: economou@env.aegean.gr 
  • Yiannis G. Matsinos
    Acoustic Ecology Laboratory, Department of Environment, University of the Aegean, Mytilene, Lesvos, Greece, email: matsinos@aegean.gr

Ecological connectivity is highly associated with conservation and could be described as the degree to which the landscape and hence the soundscape, facilitates or impedes the movement of species among resource patches at structural and functional level. Urban green areas and quiet areas offer a series of ecosystem services with a range of environmental, social and cultural opportunities for city dwellers. The specific areas could be perceived not as independent units, but as disconnected components of a landscape that require reconnection, within the urban environment. Apart from the structural characteristics of an urban landscape, environmental noise could pose as a non-physical barrier that impedes ecological connectivity at a functional level. The overall purpose of this research is to support the ecological connectivity scheme of two quiet areas in Mytilene (Lesvos Island, Greece). By means of sound recordings, several acoustic biodiversity indices, provided by the multidisciplinary field of ecoacoustics, were extracted. The noise maps created using noise prediction software highlighted the effects of noise amongst the two quiet areas. Furthermore, the visualization of the Acoustic Complexity Index inside and in between the two quiet areas highlighted the differentiation of the diversity levels and therefore the spots that are highly affected by noise at both, functional and structural level.

O62-POSTER- Soundscape and bird community at the Braulio Carrillo National Park, Costa Rica, related to highway 32.

Mónica Retamosa Izaguirre1, David Segura Sequeira2, Jimy Barrantes Madrigal1, Manuel Spínola Parallada1, and Oscar Ramírez Alán3†

1 Instituto Internacional en Conservación y Manejo de Vida Silvestre, Universidad Nacional, Costa Rica, mretamos@una.cr

2 Reserva Natural Las Arrieras

3 Escuela de Ciencias Biológicas, Universidad Nacional, Costa Rica

Natural areas are being increasingly disturbed by vehicular traffic on roads. In Costa Rica, 41.2% of protected areas are directly or indirectly influenced by roads. Many studies have associated road disturbances with noise, but the possible causes are multiple and often times intercorrelated. We characterized the soundscape and bird community in two study sites at the Braulio Carrillo National Park (BCNP) of Costa Rica, selected according to the distance to highway 32, a national road of high vehicular traffic that crosses the BCNP: “Quebrada” (contiguous to the road), and “Ceibo” (20 km away from the road). We conducted four visits to each site from June 2017 to August 2018.  We sampled the soundscape (using sound recordings), bird community (using point counts) and vegetation structure complexity (using vegetation plots and photos) in 12 sampling points in each area.  We obtained 11 acoustic indices and 4 bird indices derived from point counts. Ceibo bird community was mainly composed by forest birds, while in Quebrada we also found birds from open areas. Quebrada presented a more open forest structure and lower density of trees and shrubs; lower evenness and higher acoustic complexity, higher bioacoustic activity and sound pressure level; higher bird abundance and richness. Ceibo showed a higher density of trees and shrubs, higher complexity of vegetation structure, higher proportion of biophonies than anthrophonies, and an acoustic community with higher diversity and entropy.  Acoustic community seemed to be more diverse and better partitioned in the least disturbed site Ceibo; while less even in the most altered site Quebrada, where niches were created for open areas birds by the road effect on vegetation structure and the presence of the adjacent “Rio Sucio” river canyon. Soundscapes seemed to properly reflect the habitat condition, and are promising for evaluating the ecological condition of a site.

O63-Tuning acoustic communities: phylogenetic, functional and acoustic beta diversity

  • Manzano, M.C.R., Programa de Pós-graduação em Evolução e Diversidade, Universidade Federal do ABC, São Paulo, Brazil.
  • Sugai, L.S.M., Departamento de Ecologia, Universidad Autónoma de Madrid, Madrid, Spain. Sawaya, R.J., Centro de Ciências Naturais e Humanas, Universidade Federal do ABC, São Paulo, Brazil.
  • Llusia, D., Departamento de Ecologia, Universidad Autónoma de Madrid, Madrid, Spain.

How biophonies vary across ecological communities has been a central topic in Ecoacoustics. In each region, acoustic diversity is determined by local (alpha) and between-sites (beta) variation in sounds. While alpha diversity has been widely investigated, little is known about beta acoustic diversity. Anuran amphibians offer ideal models to tackle acoustic space-related questions and its relationship with acoustic diversity. As such, an increase in species richness is generally followed by increased local acoustic diversity. However, we still lack an understanding of how acoustic spaces vary between communities. Here, we investigated the acoustic features and composition of anuran communities in the Atlantic Forest by sampling ten different ponds in southeastern Brazil, between September 2018 and March 2019. Acoustic space can be characterized in different ways, mostly through the use of remote recorders. In this work, we explored an alternative approach based on directional recordings that estimate the occupation of frequency bins (10 classes between 650Hz and 6500Hz) for two acoustic parameters of advertisement calls: i) dominant frequency and ii) frequency ranges. We then calculated an overall beta diversity index based on the Sørensen-based multiple-site dissimilarity and assessed its turnover and nestedness components. We found a large variation in species composition and acoustic composition of communities based on dominant frequency, mainly determined by the turnover component. By evaluating communities’ pairwise similarities, we observed that similarities in dominant frequency across communities were correlated with compositional similarities, different from what was observed for the frequency range. Such findings indicate that acoustic spaces derived from “high resolution”, as dominant frequency, seem to be more efficient represent actual patterns of biological variation across communities. Still, it remains to explain the variation in cross-communities acoustic composition that are unrelated to compositional differences.

O64-POSTER- What is a Soundscape ? : New perspectives based on a functional representation

  • Grinfeder, E., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, Université des Antilles, 57 rue Cuvier, 75005 Paris, France and Equipe Audition, Laboratoire Psychologie de la Perception, Centre National de la Recherche Scientifique, Departement d’Etudes Cognitives, Ecole normale supérieure, Université Paris Sciences & Lettres, 29 Rue d’Ulm, 75005 Paris, France
  • Haupert, S., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, Université des Antilles, 57 rue Cuvier, 75005 Paris, France
  • Lorenzi, C., Equipe Audition, Laboratoire Psychologie de la Perception, Centre National de la Recherche Scientifique, Departement d’Etudes Cognitives, Ecole normale supérieure, Université Paris Sciences & Lettres, 29 Rue d’Ulm, 75005 Paris, France
  • Sueur, J., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, Université des Antilles, 57 rue Cuvier, 75005 Paris, France

The concept of soundscape was coined to describe a given acoustic environment and to highlight the complex ecosystemic mechanisms that explain its structure. Several definitions of soundscapes  have been proposed  based on musical, acoustical and ecological perspectives. However, the cause-effect mechanisms that underlie soundscapes are often overlooked. In addition, the term “soundscape” is often used in an ambiguous way, pointing to both objective and subjective realities.

Through a transdisciplinary review, we tried to identify and understand the relationships between the main biotic and abiotic factors that condition non-anthropogenic terrestrial soundscapes. We used a source-filter approach to describe sound sources, sound propagation phenomena and receiver’s characteristics. We crossed-referenced inter-disciplinary information in order to identify links between factors, sound sources and filters. We organized those relationships and the associated references in a functional diagram. Finally, we used this representation to question the different uses and meanings of the “soundscape” concept found in the literature. We then defined three separate operational notions : soundscapes, acoustic scenes and interpretations.

This study brings a new systematic approach to soundscapes that can help ecoacousticians, bioacousticians, psychoacousticians and environmental managers to better understand  soundscapes and protect natural areas in a more significant way.

P65- POSTER- An ecoacoustic study of urban forest landscape of National Capital Region of Delhi

  • Jaiswal, A., School of Life Sciences, Jawaharlal Nehru University, New Delhi, India
  • Oliveira, E.G., Department of Physiology & Behavior, Federal University of Rio Grande do Norte, Natal, Brazil
  • Sousa-Lima, R.S., Department of Physiology & Behavior, Federal University of Rio Grande do Norte, Natal, Brazil
  • Kumar, K., School of Environmental Sciences, Jawaharlal Nehru University, New Delhi, India

Ecoacoustic approach to characterize different soundscapes and its biological and non-biological components have become a widely accepted tool for monitoring habitats and its biodiversity in a particular environment. In present study, we selected five different forest sites in the National Capital Region of Delhi. We chose three different locations at each site, one at the edge of a road and others two within interior of the forest to also examine the effect of anthropogenic noise. We recorded 24h continuous ambient soundscape at these locations for consecutive five days using SWIFT autonomous recorders between April to June of 2019. We used AnalysisPrograms.exe developed by Ecoacoustics Research Group, University of Queensland to generate 24h false-colour spectrograms for visualization and to calculate different acoustic indices (AIs) for analysis. We investigated 1) how do the AIs vary across different forest sites and 2) how do they differ at locations with different anthropogenic noise conditions within each site. We examined particularly two indices namely Acoustic Complexity Index (ACI) and Normalized Difference Soundscape Index (NDSI). Our result showed that both the indices were significantly different at different forest sites showing high values for the site with high degree of native condition and minimal anthropogenic disturbance. However, within each forest site ACI values did not differ much at the three different locations but NDSI values were significantly able to reflect the varied presence of anthropogenic noise at the locations and were lower at locations close to road compared to the interior ones. The results shows the applicability of acoustic indices to characterize different habitats under different anthropogenic influences which can further be used for long term monitoring and conservation purposes.

O66-POSTER- Tropical acoustic diversity: monitoring of a seed disperser, the White throated toucan

  • Ducrettet, M., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France
  • Forget, P.-M., Mécanismes adaptatifs et Évolution (MECADEV), Muséum national d’Histoire naturelle, CNRS, 1 avenue du Petit Chateau, 91800 Brunoy, France
  • Ulloa, J. S., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France & Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Av. Circunvalar No. 16-20 Venado de Oro, Bogotá D.C., Colombia
  • Yguel, B., Centre d’Écologie et des Sciences de la Conservation (CESCO), Muséum national d’Histoire  naturelle, CNRS, Sorbonne Université, 57 rue Cuvier, 75005 Paris, France
  • Gaucher, P., Laboratoire Ecologie, Evolution, Interactions des Systèmes Amazoniens, USR 3456 CNRS,Université de Guyane, Cayenne, France
  • Princé, K., Université de Lyon, F-69000, Lyon ; Université Lyon 1; CNRS, UMR5558, Laboratoire de Biométrie et Biologie Evolutive, F-69622 Villeurbanne, France & Centre d’Écologie et des Sciences de la Conservation (CESCO), Muséum national d’Histoire  naturelle, CNRS, Sorbonne Université, 57 rue Cuvier, 75005 Paris, France
  • Haupert, S., Sorbonne Université, CNRS UMR 7371, INSERM UMR S 1146, Laboratoire d’Imagerie Biomédicale, Paris, France
  • Sueur, J., Institut de Systématique, Évolution, Biodiversité (ISYEB), Muséum national d’Histoire naturelle, CNRS, Sorbonne Université, EPHE, 57 rue Cuvier, 75005 Paris, France

Seed dispersers play a key role in the structure and dynamics of tropical forests. At the time of the Anthropocene with direct (defaunation) and indirect (habitat degradation, climate changes) impacts on wildlife, monitoring seed disperser populations is crucial for tropical conservation. Using ecoacoustics methods, we tracked in a lowland forest of French Guiana the populations of a major seed disperser, Ramphastos tucanus the White throated toucan, in a context of human pressure (hunting, logging, agriculture). We deployed an acoustic survey based on nine automatic recorders over 29 days at the beginning of the rainy season. We gathered weather data and described with GIS tools the habitat used by R. tucanus around each recorder. The vocalizations of R. tucanus were automatically identified in 22,490 recordings through a template matching method that had an area under the curve (AUC) of the Receiver Operating Characteristic curve (ROC) of 0.9184. The variation of vocal activity at spatial and temporal scale was assessed with a generalized mixed model. The automatic system detected a total of 1,748 recordings with toucans’ vocalizations. The rainfall could have a positive delayed effect on vocal activity so that R. tucanus seems to be more active before heavy rains. There is a diel pattern of R. tucanus vocalization with two peaks at 6 am and 6 pm. There was no major effect of habitat on vocal activity due to a few differences between recording sites. There was a higher vocal activity in the highest logged site compared to the agricultural one. Acoustics was a reliable strategy to monitor toucans as it revealed a clear temporal pattern and indicated a human footprint impact. This later impact could be related not only to the toucans but to the interaction within a community of seed dispersers.

O67- Acoustic indices perform better when applied at ecologically meaningful time and frequency scales”

Oliver Metcalf
Manchester Metropolitan University, Manchester, UK

Acoustic indices are increasingly employed in the analysis of soundscapes to ascertain biodiversity value. However, conflicting results and lack of consensus on best practices for their usage has hindered their application in conservation and land-use management contexts. Here we investigate whether the sensitivity of acoustic indices to ecological change and fidelity of acoustic indices to ecological communities are negatively impacted by signal masking. Signal masking can occur when acoustic responses of taxa sensitive to the effect of interest are masked by less-sensitive acoustic groups, or target taxa sonification is masked by non-target noise. By calculating acoustic indices at ecologically relevant time and frequency bins, masking effects can be reduced and the efficacy of indices increased.

We test this on a large acoustic dataset collected in Eastern Amazonia spanning a disturbance gradient of undisturbed, logged, burned, logged-and-burned and secondary forests. We calculated values for two acoustic indices: the Acoustic Complexity Index and the Bioacoustic Index, across the entire frequency spectrum (0–22.1 kHz), and four narrower subsets of the frequency spectrum; at dawn, day, dusk and night. We show that signal masking impacts the sensitivity of acoustic indices and calculating acoustic indices at a range of narrower time–frequency bins substantially increases the classification accuracy of forest classes with random forest models. Furthermore, we found signal masking led to misleading correlations, including spurious inverse correlations, between biodiversity indicator metrics and acoustic index values compared to correlations derived from manual sampling of the audio data.

Closing ceremony